---
title: "Internet Outage Statistics 2026: Frequency, Cost and Causes"
date: 2026-06-10
author: "Robert A. Lee"
featured_image: "https://sqmagazine.co.uk/wp-content/uploads/2026/05/internet-outage-statistics.jpg"
categories:
  - name: "Internet"
    url: "/internet.md"
tags:
  - name: "Statistics"
    url: "/tag/statistics.md"
---

# Internet Outage Statistics 2026: Frequency, Cost and Causes

Cloudflare detected 174 major internet outages worldwide in 2025, an average of more than three significant disruptions every week. Uptime survey data shows that more than half (54%) of organizations had a recent outage that cost more than $100,000. The data below covers outage frequency, cost, cause breakdowns, regional vulnerability, and ISP and CDN reliability.

## Key Takeaways

- [Cloudflare](https://sqmagazine.co.uk/cloudflare-statistics/) detected **174** major internet outages globally in 2025.
- **54%** of significant outages cost organizations more than $100,000, according to Uptime.
- **41%** of enterprises report hourly downtime costs above **$1 million**, per ITIC.
- **50%** of data centres had at least one impactful outage over the past three years, down from **53%** in 2024.
- Government-imposed shutdowns reached **120,095 hours** globally in 2025, a **70%** rise on the previous year.
- U.S.-centric outages peaked at **55%** between January 27 and February 16, 2025, before settling at **39%** by the end of June.
- Power remains the leading cause; IT and networking issues totalled **23%** of impactful outages in 2024.

## Editor’s Choice

- Cloudflare logged **174** major outages worldwide in 2025.
- Government shutdowns cost the world economy **$19.7 billion** in 2025.
- Russia’s shutdowns alone cost **$11.9 billion**, per Top10VPN’s tally of NetBlocks data.
- CrowdStrike’s defective Falcon update affected more than **8.5 million** [Microsoft Windows](https://sqmagazine.co.uk/microsoft-statistics/) devices.
- AWS’s October 20, 2025, outage hit **113** services for more than **15 hours**.
- Red Sea cable cut on September 6, 2025, disrupted nearly **25%** of internet traffic between Asia, Europe, and the Middle East.
- Azure Front Door’s October 29, 2025, incident lasted about **9 hours**.

## Recent Developments

- **December 30:** A major technical failure at Israeli provider Partner Communications disrupted mobile, TV, and Internet services across the country.
- **December 12:** Attacks on Ukraine damaged warehouses and energy infrastructure, with traffic dropping by as much as **57%**.
- **October 29, 2025:** An Azure Front Door configuration change disrupted Office 365, Teams, Xbox Live, and Alaska Airlines for about **9 hours**.
- **October 20, 2025:** An [AWS](https://sqmagazine.co.uk/aws-cloudwatch-statistics/) DynamoDB DNS defect cascaded to **113** services for more than **15 hours**.
- **September 6, 2025:** SMW4 and IMEWE submarine cables were cut in the Red Sea, degrading nearly **25%** of intercontinental traffic.
- **March 4, 2025:** The PEACE submarine cable was cut approximately **1,450 km** from Zafarana, Egypt, with physical repairs expected to take weeks or months.

## Internet Outage Frequency

- Cloudflare’s monitoring detected **174** major internet outages across the globe throughout 2025.
- That cadence translates to an average of more than three significant connectivity disruptions weekly.
- U.S.-centric outages peaked at **55%** from January 27 to February 16, 2025, per ThousandEyes.
- By early March 2025, the U.S.-centric share had dropped to **46%**, then **41%** by early April.
- U.S.-centric outages reached as low as **24%** during certain periods in May 2025 before settling at **39%** by the end of June.
- Uptime data shows **50%** of data centres had at least one impactful outage over the past three years, down from **53%** in 2024.
- For the fourth consecutive year, Uptime Intelligence research suggests that overall outage frequency and reported severity continue to decline.
- **The two trendlines diverge:** Facility-level outage frequency falls while internet-topology outage detection climbs. SQ Magazine reads this as failure migrating up the stack, from data-centre power to DNS, BGP, and CDN edges.

Frequency Metric2025 ValueSourceMajor global outages detected174Cloudflare RadarAverage outages per weekMore than 3Cloudflare RadarPeak U.S.-centric outage share (H1 2025)55%ThousandEyesLow U.S.-centric outage share (H1 2025)24%ThousandEyesData centres with at least one impactful outage in 3 years50%Uptime InstituteYear-over-year change in data-centre outage rateDown from 53%Uptime Institute*Source: Cloudflare Radar 2025 Year in Review, ThousandEyes, Uptime Institute Annual Outage Analysis 2025*

## The Economic Cost of Internet Outages

- More than **90%** of midsize and large enterprises report that a single hour of downtime costs their organization more than **$300,000**, per ITIC’s 2024 Hourly Cost of Downtime Survey.
- **41%** of enterprises report hourly downtime costs exceed **$1 million**, with over 90% of midsize and large enterprises seeing more than $300,000 per ITIC.
- Uptime survey data shows **54%** of organizations had a most recent significant, serious or severe outage that cost more than **$100,000**.
- One in five organizations said their most recent outage cost more than **$1 million**, per Uptime.
- Micro SMBs with fewer than **25** employees see costs estimated at **$1,670** per minute, or about **$100,000** an hour.
- **57%** of small businesses with **20-100** employees report downtime costs exceeding **$100,000** per hour, with an average cost between **$8,000** and **$25,000** per hour.
- Fortune 500 companies lost as much as **$5.4 billion** in revenues and gross profit during the CrowdStrike outage.

Cost TierShare of OrganisationsSourceOutage cost > $100,00054%Uptime InstituteOutage cost > $1 million1 in 5 (20%)Uptime InstituteHourly downtime > $300,000More than 90% (midsize/large enterprises)ITIC 2024Hourly downtime > $1 million41% (enterprises)ITIC 2024Micro SMB hourly costAbout $100,000ITIC 2024Small business average hourly cost$8,000 to $25,000ITIC 2024*Source: ITIC 2024 Hourly Cost of Downtime Survey, Uptime Institute Annual Outage Analysis 2025*

Across SQ Magazine’s [cybersecurity statistics](https://sqmagazine.co.uk/cybersecurity-statistics/) coverage, breach costs and downtime costs have tracked similar widening patterns. The hourly numbers reset purchasing arguments for redundancy, multi-region cloud, and rehearsed failover.

## Data Centre Outage Statistics (Uptime Institute)

- **50%** of data centres experienced at least one impactful outage over the past three years, down from **53%** in 2024, per Uptime.
- Outages from IT and networking issues totalled **23%** of impactful outages in 2024.
- Power remains the leading cause of impactful outages, per Uptime.
- Over the nine years Uptime has tracked publicly reported outages, third-party IT and data centre service providers have accounted for about two-thirds of those reported.
- The failure of staff to follow procedures has become an even greater cause of outages than in the previous year, per Uptime.
- Rising demand for [AI](https://sqmagazine.co.uk/artificial-intelligence-statistics/) is straining power and cooling, with growing power-stability challenges from aging grid infrastructure and intermittent renewable energy.
- For the fourth consecutive year, overall outage frequency and the general level of reported severity continue to decline.

![IT and Network Failures Driving Data Center Outages](https://sqmagazine.co.uk/wp-content/uploads/2026/05/it-and-network-failures-driving-data-center-outages.jpg "IT and Network Failures Driving Data Center Outages")

## Cloud Outage Statistics (AWS, Azure)

- On October 20, 2025, AWS experienced one of its most severe regional outages in years.
- An AWS DynamoDB update containing a critical defect triggered cascading failures that affected **113** different AWS services for more than **15 hours**.
- Affected platforms during the AWS event included Snapchat, Pinterest, Fortnite, Roblox, Reddit, Venmo, Disney+, Canva, and Amazon retail’s own support and retail systems.
- On October 29, 2025, around noon Eastern time, an unintentional configuration change within Azure Front Door triggered an outage that lasted about **9 hours**.
- The disruption impacted Office 365, Minecraft, Xbox Live, Outlook, Teams, Power Apps, Costco, Starbucks, and Alaska Airlines.
- Without DNS, AWS services could not locate their critical paired infrastructure, creating a gradual collapse during the October 20, 2025, incident.

Cloud IncidentDateDurationServices AffectedSourceAWS DynamoDB / DNS cascadeOctober 20, 2025More than 15 hours113AWS post-event coverageAzure Front Door config changeOctober 29, 2025About 9 hoursOffice 365, Teams, Xbox Live, Alaska Airlines, othersMicrosoft post-incident coverage*Source: AWS Health Dashboard summaries, Microsoft Azure post-incident review (via IncidentHub aggregation)*

For data on cloud-adjacent risk, see SQ Magazine’s [API breach statistics](https://sqmagazine.co.uk/api-security-breach-statistics/).

## Causes of Internet Outages

- Power remains the leading cause of impactful outages, per Uptime.
- IT and networking issues totalled **23%** of impactful outages in 2024, per Uptime.
- Staff failure to follow procedures became an even greater cause of outages than the previous year.
- Analysis of H1 2025 outages shows an increasing prevalence of subtle functional failures and service degradations where symptoms often appear disconnected from their root causes.
- Anchor drag accounts for roughly **30%** of global cable faults annually, per the International Cable Protection Committee analysis cited after the September 2025 Red Sea cuts.
- Cable cuts affecting both submarine and domestic fibre optic infrastructure caused service disruptions lasting from several hours to several days.
- Rising demand for AI is straining power and cooling, with aging grid infrastructure compounding the risk.

Cause CategoryShare or FrequencySourcePowerLeading causeUptime InstituteIT and networking issues23% of impactful outagesUptime InstituteStaff procedure failuresRising YoYUptime InstituteSubmarine cable anchor dragAbout 30% of global cable faults annuallyInternational Cable Protection CommitteeFunctional failures and degradationsIncreasing prevalenceThousandEyesAI-driven power and cooling strainRising risk factorUptime Institute*Source: Uptime Institute Annual Outage Analysis 2025, ThousandEyes, International Cable Protection Committee*

## Submarine Cable Cut Outages

- Submarine cables SMW4 and IMEWE were cut in the Red Sea on September 6, 2025, disrupting internet services across Asia and the Middle East.
- The cuts disrupted nearly **25%** of internet traffic between Asia, Europe, and the Middle East.
- SMW4 is operated by Tata Communications; IMEWE is managed by an Alcatel-Lucent consortium.
- The cuts resulted in degraded internet performance in India, Pakistan, and the United Arab Emirates.
- The PEACE submarine cable was cut in the Red Sea approximately **1,450 km** from Zafarana, Egypt, on March 4, 2025, disrupting internet traffic between Asia, East Africa, and Europe.
- Although cloud services were quickly restored after the PEACE cut, physical cable repairs are expected to take weeks or months.
- The International Cable Protection Committee’s early analysis points to commercial shipping activity as the probable cause of the September 6, 2025, cuts, likely a vessel dragging its anchor across the cables.

Cable EventDateCables AffectedImpactSourceRed Sea SMW4 + IMEWE cutsSeptember 6, 2025SMW4 (Tata Communications), IMEWE (Alcatel-Lucent consortium)~25% of Asia-Europe-ME trafficInternational Cable Protection Committee analysis (via NBC News)PEACE cable cutMarch 4, 2025PEACEAsia-East Africa-Europe traffic; weeks-to-months repairCloudflare Radar*Source: NBC News, Cloudflare Radar, International Cable Protection Committee*

## Government-Imposed Internet Shutdowns

- Government-imposed internet shutdowns cost the world economy **$19.7 billion** in 2025, per Top10VPN’s tally based on NetBlocks data.
- There were **212** major government-imposed internet outages across **28** countries in 2025, the highest number recorded in a single year.
- Disruptions included **120,095 hours** of internet shutdown, a **70%** rise from the previous year.
- Government-mandated shutdowns occurred in Libya, Iran, Iraq, Syria, and Panama during Q2 2025, each tied to politically sensitive events.
- Several 2025 outages were related to internet shutdowns intended to prevent cheating on academic exams, with countries including Sudan implementing regular multi-hour shutdowns over several weeks.
- Russia led with **57** shutdowns, while Myanmar and Equatorial Guinea each experienced shutdowns lasting the full year.

Shutdown Metric2025 ValueSourceTotal economic cost$19.7 billionTop10VPN / NetBlocksNumber of major shutdowns212Top10VPN / NetBlocksCountries affected28Top10VPN / NetBlocksTotal disruption hours120,095Top10VPN / NetBlocksYear-over-year hours change+70%Top10VPN / NetBlocksCountry leading shutdowns countRussia (57)Top10VPN / NetBlocks*Source: Top10VPN Cost of Internet Shutdowns 2025, NetBlocks COST methodology*

## Cost of Government Shutdowns by Country

- Russia bore the largest 2025 shutdown cost at **$11.9 billion**, per Top10VPN’s NetBlocks-based estimate.
- Venezuela incurred **$1.91 billion** in 2025 shutdown costs.
- Myanmar followed at **$1.89 billion**, with shutdowns spanning the full year.
- Russia accounted for **57** shutdowns in 2025, the highest single-country count.
- Myanmar and Equatorial Guinea each experienced shutdowns lasting the full year of 2025.

![Top Countries By Internet Shutdown Economic Loss](https://sqmagazine.co.uk/wp-content/uploads/2026/05/top-countries-by-internet-shutdown-economic-loss.jpg "Top Countries by Internet Shutdown Economic Loss")

## Most-Restricted Platforms During Outages

- X (Twitter) topped the platform-block list with **18,354 hours** blocked across 2025.
- Telegram was second with **16,990 hours** blocked.
- TikTok placed third with **14,646 hours** blocked.
- The top three blocked platforms together accounted for 49,990 hours of restriction in 2025, equivalent to nearly six years of continuous block time.
- The report draws on NetBlocks real-time data and reports, IODA, and the SFLC.IN Internet Shutdown Tracker, and OONI’s censorship measurement tools.

![Most Blocked Platforms By Hours](https://sqmagazine.co.uk/wp-content/uploads/2026/05/most-blocked-platforms-by-hours.jpg "Most Blocked Platforms by Hours")

## Major Single-Incident Outages (CrowdStrike)

- The defective software upgrade in CrowdStrike’s Falcon platform led to outages affecting more than **8.5 million** Microsoft Windows devices.
- Fortune 500 companies lost as much as **$5.4 billion** in revenues and gross profit.
- The healthcare sector recorded an estimated **$1.94 billion** in losses, the largest sector-level hit.
- The banking sector recorded an estimated **$1.15 billion** in losses.
- Fortune 500 airlines such as American and United lost a collective **$860 million**.

![CrowdStrike Outage Impact and Financial Losses](https://sqmagazine.co.uk/wp-content/uploads/2026/05/crowdstrike-outage-impact-and-financial-losses.jpg "CrowdStrike Outage Impact and Financial Losses")

## Regional Outage Vulnerability

- U.S.-centric outages fluctuated from **24%** to **55%** during the first half of 2025, per ThousandEyes.
- A massive power outage impacting Spain and Portugal disrupted connectivity within those countries in Q2 2025.
- Attacks on Ukraine on December 12 damaged warehouses and energy infrastructure, with traffic dropping by as much as **57%**.
- East Africa was repeatedly impacted by submarine cable cuts during 2025, exposing route-diversity gaps.
- India, Pakistan, and the United Arab Emirates saw degraded performance after the September 6, 2025, Red Sea cuts.
- On December 30, Israeli provider Partner Communications experienced a country-wide failure spanning mobile, TV, and Internet services.

Region2025 Outage NotesSourceUnited States24% to 55% share of global outages, H1ThousandEyesIberian PeninsulaMassive Spain-Portugal power outage Q2Cloudflare RadarUkraineUp to 57% traffic drop after Dec 12 attacksCloudflare RadarEast AfricaRepeated submarine cable cutsCloudflare RadarMiddle East / South AsiaSept 6 cable cuts, ~25% inter-continental trafficNBC NewsIsraelCountry-wide Partner Communications fail Dec 30Cloudflare Radar*Source: ThousandEyes, Cloudflare Radar 2025 Year in Review, NBC News*

## ISP and Telecom Outage Reporting (FCC NORS)

- The FCC established outage reporting rules in **2004**, per the agency’s NORS overview.
- Qualifying providers must report network outages that last at least **30 minutes** and meet specific thresholds in NORS.
- Wireline, cable, satellite, wireless, and Signalling System 7 providers must submit a NORS notification within **120 minutes** with preliminary information.
- Providers must follow up with an initial outage report within three calendar days, and a final report no later than **30 days** after discovering the outage.
- Effective April 15, 2025, service providers also comply with the FCC’s updated reporting requirements for outages that impact 911 and 988 service.

NORS RequirementThresholdSourceOutage minimum duration30 minutesFCC NORSInitial notificationWithin 120 minutesFCC NORSInitial outage reportWithin 3 calendar daysFCC NORSFinal outage reportWithin 30 daysFCC NORS911 / 988 service updatesEffective April 15, 2025FCC NORS*Source: Federal Communications Commission Network Outage Reporting System*

## Outage Patterns: Functional Failures and Degradations

- Analysis of the first half of 2025 found an increasing prevalence of subtle functional failures and service degradations.
- Symptoms in 2025 outages often appear disconnected from their root causes.
- On October 20, 2025, an AWS DynamoDB update defect cascaded into DNS failure across **113** different AWS services.
- On October 29, 2025, an unintentional configuration change within Azure Front Door rippled outward for about **9 hours**.
- SQ Magazine’s reading of the Cisco data is that operators face a detection problem more than a redundancy problem: Many 2025 incidents bypassed traditional uptime monitors because services responded slowly rather than failing outright.

PatternDescriptionSourceSubtle functional failuresSymptoms decoupled from root causeThousandEyesCascading config errorsOne change propagates broadlyAWS Oct 20, Azure Oct 29Service degradation vs outright outageSlow but live, harder to detectThousandEyes*Source: ThousandEyes Three Outage Patterns 2025, post-incident summaries*

## What Internet Outage Data Means for Businesses

- More than **90%** of midsize and large enterprises now report that a single hour of downtime costs more than **$300,000**, per ITIC.
- **41%** of enterprises report hourly downtime costs above **$1 million**, per ITIC.
- Two-thirds of publicly reported outages traced back to third-party IT and data centre service providers, per Uptime’s nine-year tracking.
- Small businesses with **20-100** employees face an average hourly cost between **$8,000** and **$25,000**.
- Resilience planning that ignores third-party concentration risk leaves the largest exposure unaddressed. SMBs should weigh provider diversity alongside pure uptime SLAs.

Business SizeTypical Hourly Downtime CostSourceMicro SMB (<25 employees)About $100,000ITIC 2024Small business (20-100 employees)$8,000 to $25,000 averageITIC 2024Midsize / large enterpriseMore than $300,000 (90%+ of respondents)ITIC 2024Top-tier enterpriseMore than $1 million (41% of respondents)ITIC 2024*Source: ITIC 2024 Hourly Cost of Downtime Survey, Uptime Institute Annual Outage Analysis 2025*

For the SMB-specific operational risk context, see SQ Magazine’s [small business breach statistics](https://sqmagazine.co.uk/small-business-cybersecurity-statistics/).

Workforce-side implications appear in our [cybersecurity workforce data](https://sqmagazine.co.uk/cybersecurity-job-statistics/) breakdown.

## Frequently Asked Questions (FAQs)

**How often do internet outages occur in 2025?**Cloudflare detected 174 major internet outages worldwide in 2025, an average of more than three significant disruptions every week. ThousandEyes data shows the U.S.-centric share of those outages ranged from 24% to 55% across the first half of the year.

 

**How much does an internet outage cost a business?**ITIC’s 2024 survey reports that more than 90% of midsize and large enterprises see hourly downtime costs above $300,000, with 41% above $1 million per hour. Uptime survey data shows 54% of organizations had a recent outage costing more than $100,000, and one in five exceeded $1 million.

 

**What are the leading causes of internet outages?**Power remains the leading cause of impactful outages, per Uptime, while IT and networking issues totalled 23% in 2024. Submarine cable damage from anchor drag accounts for roughly 30% of global cable faults annually, per the International Cable Protection Committee.

 

**Which countries had the longest internet shutdowns in 2025?**Russia led with 57 shutdowns and an $11.9 billion economic cost, while Myanmar and Equatorial Guinea each experienced shutdowns lasting the full year, per Top10VPN’s NetBlocks-based research. Total disruption reached 120,095 hours globally, a 70% rise on the previous year.

 

**What was the largest single internet outage in recent history?**The CrowdStrike Falcon update affected more than 8.5 million Microsoft Windows devices and cost Fortune 500 companies as much as $5.4 billion in revenues and gross profit. In 2025, the AWS October 20 event affected 113 services for more than 15 hours.

 

**How are submarine cable cuts affecting internet outages?**On September 6, 2025, the SMW4 and IMEWE cables were cut in the Red Sea, disrupting nearly 25% of internet traffic between Asia, Europe, and the Middle East. The PEACE cable was separately cut on March 4, 2025, around 1,450 km from Zafarana, Egypt, with physical repairs expected to take weeks or months.

 

 

## Conclusion

Cloudflare logged 174 major internet outages in 2025. Uptime survey data shows one in five organizations now sees their most recent outage cost more than $1 million. ITIC pegs hourly downtime above $1 million for 41% of enterprises. The cost surface widens at both ends of the stack, from data-centre power events to topology-level cascades like the AWS DynamoDB and Azure Front Door incidents.

Two patterns frame the year ahead. First, deliberate shutdowns now overshadow accidental outages by total economic cost, with Russia’s $11.9 billion alone exceeding many natural-cause categories. Second, failure is migrating up the stack: facility-level outages are declining, but DNS, BGP, and submarine cable events are taking their place. Operations leaders, regulators, and SMB owners benefit when planning treats both tracks as first-order risks rather than aggregating them under a single uptime SLA. SQ Magazine will continue tracking these signals as quarterly Cloudflare Radar reports and the Uptime Institute’s next annual analysis arrive.